Spatial Downscaling Of TRMM Precipitation Product Based On Precipitation Water Vapor In Mid-eastern China

The precise precipitaton product is an important basic data for flood forecasting, ecological research and climate model. However,there are some problems such as low precision and low spatial resolution, which are based on the precipitation measured and the model output. The precipitation products of remote sensing have a certain extent to solve the above problems, such as theTropical Rainfall Measuring Mission(TRMM)satellite provides high precision products for a large continuous scale and has been widely used. However, in a small range of meteorological and hydrological studies, the spatial resolution of TRMM precipitation is low. In order to overcome this shortcoming, this paper introduced the downscaling model on the basis of the water vapor factor to get the higher precision and spatial resolution of TRMM precipitation products of 1km×1km.In this paper, the study area is 98 °E-135 °E, 17 °N-50 °N .The spatial distribution of water vapor was obtained by using the Moderate-resolution Imaging Spectroradiometer Level-2 products--- MOD05 Total Precipitable Water (Water Vapor) near infrared band in Jan & Apr &Jul and Oct of 2001-2010. In order to obtain more accurate water vapor data, we joint surface and sounding data to corrected the MOD05 near infrared data. In the relationship of rainfall between geographical factors, topographic factors, vegetation factors and water vapor factors,we proposed two downscaling models by fusing multi-source data to improve the resolution of TRMM precipitation products. By comparing the error conditions of the two downscaling models results, we obtained the optimum downscaling model and obtained a higer resolution of precipitation. The main conclusions of the study are as follows:(1) MOD05 near infrared band inversion of the water vapor data with high resolution and data precision, which can describe the space distribution in complex terrain, suitabled for certain research. The accuracy of MOD05 is best in Jan and worst in Jul, Apr and Oct are placed in the middle of four typical months. Calculated by the surface vapor pressure data, we find that it is closed to the radiosonde data and MOD05 near infrared band inversion data is lower from northwest to southeast systematically. The precision of MOD05 improve significantly after revised by the fitting method and the interpolation method and the results of the interpolation method is better.(2) The results of the MOD05 downscaling model and the NDVI downscaling model are significantly better than the original TRMM data, especially in Jan. In Jul the result is in general and moderate performance in Apr and Oct. Overall,the result of MOD05 downscaling model is better than the NDVI downscaling model in Jan & Apr & Oct. And the accuracy of the result of NDVI downscaling model is slightly better than MOD05 downscaling model in Apr.(3) In the area where the precipitation conversion efficiency is low, the accuracy of the MOD05 downscaling model is slightly lower than that of the NDVI downscaling model. It is necessary to study the precipitation conversion efficiency of the local area in order to determined the downscaling model.